Conversational AI
6 min read
December 5, 2023

Great conversations are the driving force behind great hiring experiences — and generative AI is the key to unlocking them.

How do you scale things like trust, empathy, and contextual awareness? Well, if you read the title of this article then you already know.

Article Quick Links

If you’re looking for good conversation, the last place you’re probably going to search for it is a job application.

Let’s be real, the level of communication candidates typically receive as they navigate hiring processes isn’t exactly a shining beacon of quality human interaction. Most chatbots are cold and clunky, and recruiters — despite their best efforts — are usually just relegated to being the arbiters of bad news.

But here’s the thing: Even bad news can be spun into something positive in the hands (or mouth, if we’re being literal) of an elite conversationalist. 

Back when I was an HR intern many moons ago, one of my mentors was so deft at delivering dispositions that by the end of the conversation the candidates had convinced themselves they didn’t even want the job because their true passions lay elsewhere. This wasn’t some cheap Jedi mind trick — my mentor accomplished this the old fashioned way: by really listening and actually, genuinely, connecting with each person she talked to.

Since then, I’ve spent a lot of time having conversations about conversations. What makes them flourish? Falter? Fail? And within those talks, I discovered there was a lot of commonality in what people identity as the best conversations they’ve ever had:

  • Trust.
  • Empathy. 
  • Vulnerability. 
  • Shared connections. 
  • Unique perspectives.

The other similarity? They were almost all one-to-one interactions. Now none of this is really an earth-shattering revelation, I’m sure. You’ve been having conversations so long that you’re something of an expert yourself. But here’s the hard question: As it pertains to talent acquisition, how do we bottle up all those conversational qualities that work so well in close-quarter, intimate settings and scale it beyond just one candidate? 

Beyond 10?

Beyond 100?

Beyond 1,000?

Generative AI cracks the code on great conversations at a scale.

You’re probably expecting me to say that the answer is generative AI.

And I am. But I’m also not.

As is the case with any and all TA-related tasks, we can’t just rely on AI. It’s not a magic wand we can simply wave over hiring processes for instant success. I firmly believe that the secret sauce here is — and always will be — a balanced blend of AI automation and scalability combined with human judgment and ingenuity. 

At the end of the day, we want actual people (recruiters and hiring managers) to own the critical conversations with candidates. That’s where the magic actually happens, and that’s how you ensure you’re hiring the best candidate for each role at your company. But the reality is that it’s quite literally impossible for a recruiter to have focused conversations with dozens of candidates (let alone hundreds) in a single day, at every step of the hiring process.  

That’s where generative AI comes in.

It allows you to have truly great conversations with all candidates for all roles at any time of the day and during any point of the hiring process. 

That means your candidates are much happier. Your conversion rates go way up. Your turnover goes way go. And your recruiters have more time to invest back into delivering great conversations themselves. 

Here’s what that looks like. 

How generative AI establishes trust and builds empathy with candidates. 

Most chatbots are not smart. Or helpful. And because of those (fairly accurate) preconceptions, the very first thing candidates are trying to figure out when they engage with one is can this thing actually help me, or is it dumb?

Fair or unfair, people are bringing in ten pounds of bad chatbot baggage into your application process — if a candidate begins to engage with a chat widget on your career site or is going through a text to apply process on their phone, your AI has about one second to prove your usefulness to them before the ghost of all those awful past experiences scares them away.

The best thing about generative AI is that it’s much more competent in providing value than chatbots of yore. 

Competence = trust. 

Here’s an example of what our conversational assistant (who we call Olivia) can do now that we’ve layered on generative AI: 

Let’s say that a candidate tries to really put Olivia’s competence to the test by giving a fake name like, let’s say … Boaty McBoatface. Generative AI helps Olivia “learn” and “memorize” over time, allowing her to recognize that this absolutely isn’t real name, and respond in kind by saying “Eh, probably not. Why don’t you give me your real name so I can better assist you?”

This ability to learn also allows Olivia to have more situational awareness and be more emphatic during crucial moments. Most chatbots can only answer the simplest questions in the simplest ways, void of all of the nuance and emotional intelligence we’re used to when we show vulnerability around actual people. If your house was on fire and you called 9-11 asking to send help, the person on the other end isn’t going to just say “OK!’ and hang up. 

There’s probably going to be some follow up, right?

With generative AI, Olivia can now answer the complicated questions in addition to the simple ones, right in the same response:

Olivia sending that candidate a list of ways to prepare for their interview is good. Being able to layer in more empathy and reassurance makes it great.

How generative AI can handle multiple and complex intents.

That layer of empathy is incredible — but an important thing to point out is that most chatbots wouldn’t even be able to answer a second question at all, let alone with emotional intelligence.

Most current chatbots subscribe to a “one and done” model. Two questions at once? Does not compute.

This is a problem, because that’s not how humans talk. Great conversations are complex, and great conversationalists know how to nimbly navigate that complexity. I learned early on in my career that a lot of candidates like to share their entire life’s story in the first paragraph they send, and we previously had no way of addressing this in a way that felt natural or helpful.

It’s hard to overstate what a profound leap forward this is.

With generative AI, Olivia can not only identify the right intent when a complex question is asked (which she couldn’t do before), she can now identify multiple intents and craft a specific response to each of them.

This gives her infinite opportunities to add value and build shared connections with candidates. Can I bring my bulldog to work? Do you guys celebrate Taco Tuesday? Will your health insurance cover pickleball injuries?

Generative AI enables Olivia to answer everything. 

How generative AI can add context during the candidate journey.

What else?

No, really. Olivia can now answer that, too. 

We’ve talked about complex intents, but another wrinkle is that generative AI enhances Olivia’s contextual awareness as well. If someone asks a follow up question that, if isolated, makes no sense, Olivia has the sophistication to pick up on it. It’s the “what else”, so to speak:

This goes for more temporal aspects of the candidate experience as well. If a candidate already knows us, then Olivia knows them: Hey, I saw you applied on LinkedIn a month ago and I thought you might be interested in this new open role!

There’s nothing worse than when you’ve been a loyal customer of a certain brand for years and lazy automated messaging treats you like a stranger. It doesn’t feel good. It feels even worse when you’re a candidate applying for a job you might spend the majority of your life at for the foreseeable future.

Context matters. 

Generative AI is the new gold standard for candidate experience.

Look, people have been telling me not to automate stuff in the hiring process for a long time. 

“You can’t automate that. You need a person there.”

Do we?

Do we really need one-to-one human interaction at every stage of the process, or do candidates just want things to be fast and right?

Think about this: If you’re at a grocery store grabbing a few items, do you prefer the self-checkout or being rung up by a person? No offense to any grocery cashiers (they’re doing great work!), but I don’t think most of us are getting any self-actualization from small talk with them. We just want to buy our sparkling water and granola bars and get the heck out of there.

The same thing goes for certain parts of the hiring process. Does a candidate want to talk to a person to get scheduled for a phone screen, or do they just want that portion of the process to be fast and right? We have tons of data and real life experience that says it’s the latter. 

Here’s the great news: With generative AI it can be both.

Fast and right, and an actual great conversation.

Something we say about Olivia quite a bit is that we want candidates to consciously know she’s not a real person, but subconsciously think she is. We’re not trying to trick anyone, but we literally still have thousands and thousands of candidates who talk to her like she’s a friend, not an AI. The amount of times someone says “thank you” to her or sends her a smile emoji is genuinely staggering.

And that was before generative AI. I’m excited to say that this new version of Olivia is the current gold standard for AI in talent acquisition. 

So now if you’re looking for a great conversation, you can avoid grocery store checkout lanes or community meet-and-greets — and just apply for a job instead.

Written by
Adam Godson
President & Chief Product Officer
Adam Godson
Written by
Back to top

Every great hire starts with a conversation.

Demo Olivia now